Primary language, version 3: ‘What is your main language?’

Following concerns that language spoken at home may not be clear to respondents ONS tested an alternative wording, asking ‘What is your main language?’ in the May 2008 module of the Opinions Survey as well as subsequent postal tests and waves of cognitive testing (figure 6.21).

In the second part of Wave 6 ONS changed ‘Other, write in (including sign language)’ to the plural ‘Other, write in (including sign languages)’ to include all possible sign languages and not just one sign language, which may have been confusing (Figure 6.22). Concerns that respondents would write in ‘sign languages’ rather than a specific sign language led ONS to change the instruction to ‘including British Sign Language’ for the March 2009 and July 2009 postal tests (Figure 6.24). For the English version of the question in Wales, ONS tested both the wording ‘English or Welsh’ (Figure 6.22) and ‘Welsh or English’ (Figure 6.23) for the first tick-box.

Figure 6.22: Primary language question version 3 with wording ‘English or Welsh’ (Wales, English)

Figure 6.23: Primary language question version 3 with wording ‘Welsh or English’ (Wales, English)

language)

Figure 6.24: Primary language question with ‘including British Sign Language’

One disadvantage of this question is that it does not collect the full range of languages spoken as respondents will only enter a language other than English if it is their main language. While this means that some information on language diversity will be lost, other language questions tested have been found to produce unreliable data on language diversity due to inconsistent choices over whether to include basic abilities, such as a language GCSE.

Main language was considered a useful concept in meeting the essential user need of allowing data users to understand which languages services should be provided in. In order to best meet this need, and to give an indication of language diversity, it is important that respondents report their first or preferred language rather than ticking ‘English’ because, for example, they speak it more often. It was therefore important to determine whether respondent understanding of the term ‘main’ language was consistent with the intended meaning of ‘first’ or ‘preferred’ language rather than meanings relating to frequency of use or current usage. Table 6.25 shows the results of the June 2008 Opinions Survey for a question that was included asking ‘which of these is true of your main language?’ The question encouraged multiple responses, so the total percentage of choices is greater than 100.

Table 6.25: Response percentages for multiple response question ‘which of these is true of your

main language?’, Opinions Survey June 2008

Response Percentage of cases

It is the first language I could ever speak 95.9

I use it most of the time 56.9

I speak it when I am at home 58.0

I speak it with my family 58.0

I speak it with my friends 57.7

I speak it at school or work 48.1

I use it when dealing with public authorities 54.1

Total 428.8

Table 6.25 shows that 96 per cent of respondents felt that their main language was the first that they could ever speak. There was significantly less agreement that the other concepts were true of main language. This pattern was more exaggerated amongst those who had previously answered that their main language was not English and suggested that they typically saw main language as referring to the first language which they spoke and the language they used at home and with family.

This appears to be supported by cognitive testing. Many respondents with English as their second language who used English on a daily basis interpreted the question to be asking for ‘mother tongue’ or ‘first’ language and recorded the other language they spoke. Respondents often equated this with the language that they used at home with their family or the one that they felt most comfortable with and were able to express their thoughts in:

‘Your first language, the first language you speak … the one you were taught as a child … what your family taught you.’ (Wave 6)

‘The language which I speak at home and every day, and which I think in.’ (Wave 6) A small number of respondents interpreted ‘main’ language in cognitive testing as the language spoken most often:

‘The one that you use most.’ (Wave 6)

However the majority of respondents interpreting the question in this way had English as a first language so the language they spoke most often was their mother tongue. Their answer was therefore consistent with the answers of those interpreting it as ‘mother’ tongue.

One respondent whose mother tongue was Afrikaans understood the question to be asking about frequency of use rather than mother tongue and answered ‘English’ as their main language. Although such answers would reduce the quality of information on language diversity, the information provided would still meet essential user needs appropriately as anyone who mainly used English as their primary mode of communication would have appropriate language skills to be able to access services adequately in English.

Table 6.26 shows response percentages for the main language question in the 2008 Postal Test and March 2009 Postal Test. Response percentages for the two tests were very similar. The question returned a low non-response rate of around 2 per cent in both tests and only 0.5 per cent of respondents chose to tick more than one box. In both tests around 2.5 per cent of respondents indicated a main language other than English, suggesting that the question does collect

Table 6.26: Response percentages for the main language question in the 2008 Postal Test and

March 2009 Postal Test

Response 2008 Postal Test March 2009 Postal Test

English 95.1 95.1

Other 2.4 2.6

Ticked both boxes 0.5 0.5

Non response 2.0 1.9

Total 100 100

Recent Welsh language Whole Questionnaire Testing found that although most of the respondents answered the main language question correctly, many of the UK born respondents were confused by the way this question is worded, and expected to have to make a choice between ‘Welsh’ or ‘English’ as their main language. One respondent was particularly angered by the question, and crossed out English in the response option. Another respondent refused to answer this question, because he felt strongly that there was no such thing as a ‘main’ language.

The Welsh language Whole Questionnaire Testing recommendation for this question is that there should be a publicity campaign for Welsh speakers in Wales to raise awareness about the purpose of this question to minimise any confusion.

In Wave 6 of cognitive testing, ONS removed the routing and changed the order of ‘Welsh’ and ‘English’ in the first tick-box on the English language version for Wales (figure 6.23). Details of the testing conducted to determine whether to include routing is presented in section 6.3.2 below. There were no strong opinions expressed by respondents in Wave 6 about which language should be listed first. Quotes in relation to the ‘Welsh or English’ tick-box include:

‘I didn’t think a lot about it … does it matter? Will people complain? It doesn’t make any difference to me.’ (Wave 6)

‘The Welsh will be happy, grand, no problem … it’s normally English or Welsh, so it’s normally the other way round’. (Wave 6)

Some Welsh-speaking respondents were confused by the question. A few thought initially that the question was asking ‘which of these is your main language, “Welsh” or “English”?’ and were confused. Others were glad that they were not asked to make the choice between ‘Welsh’ or ‘English’ as they felt that they were bilingual and that both languages were their ‘main languages’. To minimise this type of confusion, the language in which the questionnaire is written will appear first. The English language version of the questionnaire in Wales will therefore read ‘English or Welsh’ and the Welsh language version will read ‘Cymraeg neu Saesneg’ (‘Welsh or English’). Although the findings suggest there is potential for some confusion in Wales about the ‘Welsh or English’ tick-box, particularly for Welsh speakers, version 3 does not suffer from some of the problems encountered in previous versions. The question is easily understood by the majority of respondents and understanding of the term ‘main language’ was found to be comparatively consistent amongst respondents. Unlike other questions considered, testing has suggested that those with only basic ability in a language will not distort the data by writing in the language. The question also fits within the space available on the census questionnaire. For these reasons, version 3 is the recommended primary language question for the 2011 Census.